Can Histology Be Predicted By Radiomics From CT Characterisation In Small Renal Masses (<4cm)?

Abstract

Objective- The objective of this study was to evaluate the diagnostic capability of new technology which involves the utilization of texture analysis as a radiomic marker for the differentiation of renal neoplasms.

Materials and Methods- The study involved 32 patients who had been pathologically assessed to have one abnormal kidney, with the other physiologically normal. The texture characteristics were derived from both the neoplastic endowed kidney and the normal kidney, through analysis of the CT images produced. Measures of central tendency and measures of dispersion were employed by using ImageJ software to analyze the mean, standard deviation, mode, median, skewness and kurtosis.

Results- From the 32 samples’ characteristics, there was a notable variation between the means of the normal tissue and the cancerous tissue of individual patients. It is also imperative to note that there were slight variations in the other parameters used.

The standard error of the mean was deduced to be 32.802 and the p value as <0.00001 with respect to significance levels 1%, 5% and 10%. Conclusion- Texture analysis embedded in CT imagery can be practically employed in actual noninvasive radionomic contrasting of the different forms of renal neoplasms.

Key Points • However, much it can be employed for neoplastic diagnosis cum differentiation, CTTA is not immune to drawbacks. • Texture analysis enables medical personnel to differentiate between different renal tumors. • Texture analysis when combined with pathomics is quite reliable and thus patients could evade unnecessary invasive diagnostic and therapeutic procedures.